AI Paper Atlas

Research map / Graph learning / Graph self-supervised learning

MoE-Guided Graph Diffusion for Oriented Molecule Design

Shuochen Li, Xiangqi Guo, Huobin Tan, Lei Shi

AAAI 2026 · Graph learning › Graph self-supervised learning · approach: Molecular graph pretraining · 0 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add RL finetuning after MoE-guided pretraining (ablation)MEGD · QM9 full-sampleGap: 7.107 -> 9.110Improved
add RL finetuning (ablation)MEGD · ZINC250k full-sampleGap: 5.465 -> 7.859Improved
add timestep input to the MoE gating network (ablation)MEGD · three target properties after RL finetuningperformance: 17.8% drop when removedImproved
add sub-structure input to the MoE gating network (ablation)MEGD · three target properties after RL finetuningperformance: 26.1% drop when removedImproved
use both timestep and sub-structure gating inputs (ablation)MEGD · three target properties after RL finetuningperformance: 41.4% drop when both removedImproved
replace GDM layers with MoE layers (pretraining only) (ablation)graph diffusion model · QMOFvalidity (%): 63.669 -> 72.315Improved
replace GDM layers with MoE layers (pretraining only) (ablation)graph diffusion model · QM9validity (%): 99.498 -> 99.220No clear effect
select two experts instead of one (ablation)MEGD · hyperparameter tuning, 6 total experts, balance weight 0.01target property score: 9.069 -> 9.926Improved
increase total experts from 6 to 8 (two selected) (ablation)MEGD · hyperparameter tuningtarget property score: 9.926 -> 9.080 at balance weight 0.01Hurt
use a large load-balance weight (1) instead of moderate (ablation)MEGD · hyperparameter tuning, 2 of 8 expertstarget property score: 9.080 -> 8.357Hurt
MEGD vs best RL-guided diffusion baseline GDPOMEGD · QM9, ZINC250k, QMOF full-sampleGap: 8.110 -> 9.110 (QM9)Improved
MEGD vs MOOD on top-25% ZINC250k samplesMEGD · ZINC250k top-sampleGap: 11.361 vs 11.296Hurt

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy